Github user EronWright commented on the issue:
https://github.com/apache/flink/pull/4636
In certain environments today, the Hadoop dependencies are actually
duplicated on the classpath, i.e. in the Flink dist jar and also via the local
installed Hadoop (when something like `HADOOP_HOME` is set). After this
change, will having Hadoop on the classpath (not in the dist jar) be
sufficient? I think so, just confirming.---
